Better Discussions Raised Group Intelligence in Studies

posted by jmorgan on Tuesday, March 22, 2011
jmorgan's picture

The existence of a trait we call "intelligence" is well-proven for individuals. It can be measured, relates to performance on a range of tasks, and predicts future performance. Studies into group performance have sometimes compared the average intelligence numbers for all team members, or the intelligence of the smartest member, to group performance. But could there be a "collective intelligence" separate from those measures, and does it affect performance?

A team of business, psychology, and communications researchers tested the idea through a pair of studies. They created groups of two to five people recruited from the public and had them complete general tasks such as "solving visual puzzles, brainstorming, making collective moral judgements, and negotiating over limited resources." In one study they also played checkers as a group against a computer, and in the other performed a more complex "architectural design" task.

How well the team did on one general test was linked to how well it did on others, and the score on those tests predicted how well it did on the more complex task. However, neither the average intelligence of the members or the score of the smartest member were related to the group's success. What the researchers called the "c factor" was something else, and it explained nearly 45% of the difference between groups in their performances.

Average and highest member intelligence was modestly correlated to collective intelligence, but these were not the strongest factors. Further tests found that out of "a number of group and individual factors that might be good predictors of c," only three were: social sensitivity, specifically the ability to read someone's emotions through their eyes; how many and how much people joined in the group's conversations; and the percentage of females in the group. However, the authors point out that females typically rate higher on social sensitivity, so those factors are related. Other possibilities you might expect to relate to performance, "such as group cohesion, motivation, and satisfaction" or personality, were not related to collective intelligence. Higher involvement in discussions by more people was a bigger part of the c factor than were average or highest individual intelligence, with sensitivity less related.

As the scientists suggest, "it would seem much easier to raise the intelligence of a group than an individual," using better collaboration tools, for example. (Editor's note: In fact, the results appear only to have identified that better discussions have a big impact on performance, as detailed in my related blog post, "'Group Intelligence,' or Simply Better Discussions?").

Sources:

  • Share/Save